In 1989, Fraas and Avery demonstrated a world-record 31% efficient AMO GaAs/GaSb tandem solar cell. This record efficiency still holds today. However, the GaAs/GaSb mechanical-stacked cell was designed to work with concentrated sunlight and at that time, the space community had no experience with concentrated sunlight solar arrays. So the photovoltaic community continued to work on improving flat plate cell efficiencies for satellite power systems. This work led to the adoption of the InGaP/GaAs/Ge monolithic tandem cell with an efficiency of 23%. Meanwhile in 1992, Fraas and Avery fabricated GaAs/GaSb cells and Entech supplied lenses for a concentrator min-module that was flown on the Photovoltaic Advanced Space Power (PASP) satellite. This mini-module performed well with high power density, excellent radiation resistance, and absolutely no problems with sun tracking. The success of the PASP module then led to the successful use of a 2.5 kW line-focus concentrator array as the main power source on Deep Space I. Deep Space I was launched in 1998.
So now, ten years later, concentrating solar photovoltaic arrays are proven. However, Deep Space I used the 23% efficient InGaP/GaAs/Ge monolithic cell in spite of the fact that the mechanically stacked GaAs/GaSb cell had a higher efficiency. This observation leads us to a discussion of the relative trades of monolithic vs. mechanically stacked cells.
Both monolithic and mechanically stacked tandem cells achieve high efficiencies by using multiple materials with multiple junctions sensitive to different portions of the sun""s spectrum. In the case of monolithic cells, the different materials are grown as thin single crystal films on a single substrate. This approach is lighter and more elegant. It leads automatically to a two-terminal device that looks just like a simple traditional single-junction cell. Because of its light weight, this approach is preferred for planar space arrays. However, the constraint that the layers are grown as single crystal layers in succession on the same substrate is quite restrictive. All the layers need to have a common crystal structure with a matched atomic spacing. Furthermore, because the multiple junctions are automatically series connected, the junction with the lowest current limits the current of all of the other junctions. It is therefore desirable to match the currents produced in each junction to maximize efficiency. Additionally, manufacturing yields are lower with monolithic cells because of their built in complexity.
Mechanically stacked cells involve separate junctions grown on separate single crystal substrates. This approach does not have the constraints on crystal structure, crystal atomic spacing, or current matching in series connected junctions. To date, the only successful monolithic cell is the InGaP/GaAs/Ge cell, and the only successful mechanically stacked cell is the GaAs/GaSb cell. The atomic spacing for the GaAs and GaSb cells are not matched, but they need not be for a mechanical stack.
The InGaP, GaAs, and Ge crystals, structures and atomic distances are matched. In the dual junction form with junctions in the InGaP and GaAs layers, the currents are matched at 16 mA/cm2 at I sun illumination. However, the dual junction GaAs/GaSb cell outperforms the dual junction InGaP/GaAs cell because the GaSb cell captures solar energy not available to the InGaP/GaAs cell. In other words, the GaAs/GaSb pair responds to the sun""s spectrum over the range between 0.4 and 1.8 microns, whereas the InGaP/GaAs cell only responds to the spectral range between 0.4 and 0.9 microns. This larger response range explains why the GaAs/GaSb cell outperforms the InGaP/GaAs dual junction cell.
In recent times, the manufacturers of the monolithic cells have made some improvements by placing a third junction in the Ge substrate. This approach has led to an efficiency of 27%. However, this approach has still not closed the gap with the GaAs/GaSb mechanically stacked cell. There are two reasons for this. While the Ge cell can receive energy in the spectral range between 0.9 and 1.8 microns, the Ge cell is not current matched with the InGaP/GaAs pair. Its excess current cannot be used. The current from the GaSb cell an the other hand is not limited. Furthermore, the Gasb cell generates 20% more voltage. Additionally, the Ge cell performance degrades rapidly with high-energy particle irradiation.
So, there are good reasons to work with mechanically stacked cells. However, there has been a barrier. Referring to FIG. 1, the traditional mechanically stacked cell is a four terminal device with + and xe2x88x92 connections on both the top and bottom cells. However, the user community wants two terminal devices. While we have previously described two terminal voltage matched circuits as a solution to this problem, in practice, circuit assembly to date has been very complex. A principal reason for this complexity has been the contact to the bottom of the top cell. This backside contact pad is not readily accessible.
There is a need for a two terminal circuit using mechanically stacked cells where the circuit assembly can be accomplished inexpensively using conventional automated circuit assembly equipment. There is also a need to increase the efficiency for these circuits to still higher values.
The object of the present invention is a two-terminal cell-interconnected circuit consisting of a substrate with bonding pads for bonding a first set of infrared sensitive cells with a second set of visible light sensitive cells mechanically stacked and adhesive bonded on top of the first set of cells. The substrate contains metal traces for wiring the bottom cells in series. It also contains metal traces for wiring the top cells in parallel. The number of series connected bottom cells and the number of parallel connected top cells are chosen such that the voltages produced by the top and bottom cell circuits at maximum power are nearly equal. A unique feature of these circuits that makes them easy to manufacture is that the + and xe2x88x92 contacts for the top cells are both on the topside of the top cells. Herein, we refer to these circuits as mechanically stacked cell-interconnected-circuits (MS-CiCs).
In these circuits, the top cells are either single-junction AlGaAs/GaAs cells or dual junction InGaP/GaAs cells (grown on IR transparent GaAs substrates). In either case, these cells respond to sunlight in the spectral range between 0.4 and 0.9 microns. The bottom cells can either be GaSb cells or InGaAs/InP cells. Again, in either case, these bottom cells respond to the sun""s spectrum between 0.9 and 1.8 microns. The top cells are fabricated with a grid on their backsides so that they allow the spectral power between 0.9 and 1.8 microns to pass through to the bottom cells.
Several companies have reported on InGaP/GaAs cells with the highest reported efficiency being 26.9%. Given InGaP/GaAs top cells, JX Crystals has recently fabricated four terminal stacks using GaSb bottom cells where the GaSb cell boosted the efficiency of the top cell by 6.5 percentage points. This implies that an efficiency of 26.9%+6.5%=33.4% is achievable.
These and further and other objects and features of the invention are apparent in the disclosure, which includes the above and ongoing written specification, with the claims and the drawings.